UMI if

INFORMATION TO USERS

Thismanuscripthasbeenreproduced from the microfilm master. UMI filmsthe textdirectlyfrom the original or copy submitted. Thus, some thesisanddissertation copies areintypewriterface.while others may be from anytypeofcomputer printer.

The quality oftbis reprodactioD is depeadeot upon the quality oftbe copy submitted. Broken orindistinctprint, colored or poor quality illustrations and photographs. print bleedtbrough, substandardmargins, andimproper aligmnent can adverselyaffectreproduction.

Inthe unlikely eventthatthe author did not send UMI a complete manuscript and there aremissingpages, thesewillbe noted. Also,if unauthorized copyright materialhadtoberemoved, a notcwillindicate the deletion.

Oversize materials (e.g., maps,drawings,chans)are reproducedby sectioning the original,beginningat the upper left-hand comer and continuing fromleft torightinequal sectionswithsmalloverlaps. Each original isalsophotographedinonc exposureandis includedinreduced form at the back ofthe book.

Photographsincluded inthe original manuscript have been reproduced xerographicallyinthis copy. Higher quality6"x9"black and white photographic prints are available for any photographs or illustrations appearinginthiscopy for an additional charge. Contact UMI directly to order.

UMI

ABell&:'HQWl:UlnforrnationCo!llpany 300NonhZeebRoaa,AnnA1bor MI 48106-Il46USA

313n61-4700 8OOIS21~

FOOD AND FEEDING REQumEMENTS OF JlNENILE STRIPED WOLF'FISH (Anarhichas lupus).

by

Sherra D. Fam

A thesis submitted10the School of Graduate Studies inpartialfulfilment of the requirements for the degree of

Master of Science.

AquacultureUnit,School of Fisheries Fisheries and Marine [nstitute of MemorialUlliversityofNewfoWldland

Febnwy, 1997

St.John's NewfoWldland

1+1

NalionalLibr81)' of Canada

Bibliotheque nationale du Canada Acquisitions and Acquisitio~et BibliographicServices services blbliographiques 395W"'irIglonStrwt 395... Wellin9lon

= 0 0 K1AQN4 = O N l(1ADN4

The author has granted a non·

exclusive licence allowing the NationalLibraryof Canada to reproduce, loan. distnbute or sell copies of this thesisinmicroform., paper or electronic formats.

The author retains ownership of the copyrightinthis thesis. Neither the thesis nor substantial extracts fromit may be printed or otherwise reproduced without the author's permission.

L'auteur aaccordeune licence nOD exclusive permettant

a

1a Bibliotheque natiooate du Canada de reproduire, prater, distribuer au vendre des copies de cene thesesallS 1a forme de microfiche/film, de reproduction sur papier au sur format electronique.

L'auteur conservela propri6te du droit d'auteurquiprotege cette these.

Ni la thesenides extraits substantiels de celle-ci ne doiventetre imprimes au autrement reproduits sans son autorisation.

0-612-23131-3

Canada

Abstract

The striped wolffish.(Anarhichas lupus) is a candidate species for commercial aquaculture in NewfoWldland. It possesses a number of characteristics which facilitate culture, such. as large eggs, well developed larvae wbich. readily accept formulated feeds and tolerance to low temperatures. Little research bas been conducted10determine the dietary requirements of the juvenile wolffisb or the optimum slocking density and feeding frequency.

The effects oftbree feeding frequencies (two meals/day, one meaVday and one mealltwo days) on variousgrowthparameters were investigated.Meanmeal sizewassignificantly and inversely affected by the feeding frequency.Inaddition, total feed consumption over time was directly affected by the meal frequency. The spedfic growth rate (SGR)was not adversely affectedby the decrease in meal frequency or feed intake.Feedand labour costs, therefore, may both be reduced by lowering the frequency, without compromising theSGR.

The stocking density also affected the feed consumption. The smallest mean meal size (3.990 mgfg fish)was conswned by fish stocked at 80gIL. The largest meals were consumed by fisb stocked at 50gIL(4.955 mgfg) while the meal size of fish Slocked at 20 gILwasin between these values. The feed convernion ratio (FeR)decreased significantly whentheSlocking densitywasgreater !han SOgILand !he protein efficiency ratio (PER)wassignificantly higher when the stocking densitywasgreaterthanSOgil.

The dietary energy balance, expressed as the protein energy:tola1 energy ratio (pE:TE) had a significant negative influence on the intake of feed, lipid and energy. As the PE:TE increased, the feed., lipidandenergy intakesalldecrease significantly. The PE:TE had no significant impact on the FCR or PER at either constant 9 °C or ambient temperatures (13.0

°c

to 2.00c). The production cost (based on feed costs per kilogram offish produced) was not significantly affected by !he PE:TE orbydecreasing temperalureS.

Acknowledgements

Many people and organisations have been instrumentalinthe completion of this research.

The fmancial assistance of the Canadian Centre: for Fisheries Innovation is gratefully acknowledged. The Fisheries and Marine Institute of Memorial University of NewfoWldland provided use of their recirculaled seawater facility and laboratories.

Valuable hatchery space and technical assistance were generously supplied by John Watkins and the staffofthe Wesleyville Marine Fish Hatchery.

I am grateful for the support provided by my supervisory comminee, Dr. Slephen Goddard (Fisheries and Marine Institute) andDr.Joe Brown (Ocean Sciences Centre).

Their input throughout the experimenu and their critical reviews of the manuscript were invaluable. My sincere thanks go coDr. Jay Parsons (Fisheries and Marine Institute) for his assistancewiththe statistical analysis of my data and for his editorial contributions.

Several other people have given of their time and skills to assist me throughout this research. Ray FilZgera1d,the technician in the St. John's lab facility, provided many hours of practical assistance. I am gratefuJ tohimnoc only for his kind help, but for making long hours in the tab fly by.

r

am also indebted 10 Dinah Kirby, Keith Mercer, Stephanie Power and David Yanchus, all of whom aided me in their own special ways.

I am only one of several people who have been studying wolffishinNewfoundland. I am pleased to have been able 10 work alongside Dena Wiseman, Kelly Moret, Laura HalfYard andJohn Watkins. The network of shared information and resources they helped to create has been most valuable. Ithankthem for their inputand enthusiasm.

Iamtruly grateful for the immeasurable support [ have received from my loving husband Ham. who didn't mind sharing my attention with a thousand WQlffish during our courtship. My family and Hani's parents have been wonderfuUy curious about my research since the very beginning, and are certainly as thrilled about itS completion as I am.Their faithfulness in prayerhasbeen a constant source of encouragement. Above all, thanks be to God by whose grace this work is complete.

Table of Contents

ABSTRACl' ACKNOWLEDGMENTS UST OF TABLES LIST OF FIGURES LIST OF APPENDICES 1.0 lNTROOUCTION

I.lBiological Adaptations 1.1.2TemperarureTolerance

1.1.2.1 Eggs andLarvae:

1.1.2.2 Juveniles and Adults l.lo3Fecundity

1.1.4 Egg Size and Larval Development 1.1.5 Growth

1.2Culture Methods 1.2.1StockingDensity 1.2.2Diet 1.203Protein Requirements 1.2.4 PE:TERequirements 1.2.4.1Hepatosomatic Index 1.2.5 Feeding Frequency 1.3Rationale

2.0MATERIALS ANDMEmODS 2.1 Aquarium Facilities

2.1.1RecircuJationFacility 2.1.2 Flow-ThroughFacility 2.2 Fish

2.3Esperimental DesigD

203.1Feeding Schedule/StockingDensityTrial 203.2 Dietary Energy Trial

2.3.3 Gastric Evacuation Trial 2.3.3.1 Temperature Study 2.3.3.2 Dietary Energy Balance Study

iii viii

xii 1 1 1 2 2 J Error! Bookmark not derIDed.

6 6 7 7 8 9 [0 [0 11 13 13 13 17 17 '0 20 20 21 21 22

2.4FtedJ

2.4.1 DensitylFeedingScheduicTrial 2.4.2 Dietary EncrgyTrial 2.SFftding Protocol 2.6Sampling

2.6.1 Wcighing 2.6.2 Gastric Evacuation Trials 2.6.3 Dict Analysis

2.'

CalcultotiolU 2.7.1 Condition[ndc"

2.7.2 SpecificGrowthRate 2.7.3 Feed Coovenion Ratio 2.7.4ProteinEfficiency RAtio 2.7.5HCpalOSOmatic Index 2.7.6Cost of Production 2.8Statistic.al Analy5Lt

2.8.1Feeding Trials 2.8.2 Gastric Evacuation Trials 3.0 RESULTS

3.1 Stoc:kiDg DelUityl F«di.oa: Scbedule Trial 3.1.1Observations

3.1.2MotpbometriC1 3.1.3 SpecificGrowthRate 3.1.4FeedConsumption 11.5 Feed Conversion Ratio 3.1.6ProteinEfficiency RAtio 3.2 Dietary Eberg)' B....ce Trial

3.2.1Temperarures 3.2.2Morphometries 3.2.3 SpecificGrowthRate 3.2.4 Feed Consumption 3.2.5 DailyLipidIntake 12.6 DailyEnergyIntake 3.2.7HepalO50matiClodex

22 22 22 23 24 24 25 26 27 27 27 28 28 28 29 2.

29 30 31 31 31 J2 J2 36

3.

38 38 38 40 49 52 55 58 61

3.2.8 Feed Conversion Ratio 3.2.9 Protein Efficiency Ratio ].2.10 Production Cost 3.3. Gastric [ncuatioD Trial

].].1 PE:TE Trial ].].2 Temperature Trial 4.0 DISCUSSION

4.1 Srodual DelUiry 4.2 F«diDg Schedule

..t.)P£:TE 4.4 Gutric EmptyiDg 4.5 Btparo,omaticIadn 4.6AquacultureImplic:atiolU 5.0CONCLUSIONS

5.1 Stocking DtD5ity' FtediDl Schedult Trial 5.2 Ditrary Entr-gy aalanct Trial 5.3 GutricEvacuatioDTriall REFERENCES

PERSONAL COMMUNICAnONS APPENDICES

vii

64

.,

64 72 72 72 78 78 79 80 81 8J 8S

..

8.

8' 8'

'"

101 102

List of Tables

Page Table 1.1: Comparison ofeggdiameter and body lengths at Ilatch of several.

cold-Water marinefishspecies.

Table 2.1: Ingredients and proximate analysis ofexperimental diets. 19 Table 3.1: lnitial and final lengthsandweights ofwolffish. A: stoeking density 33

mal (mean±standarderror); B: feeding schedule mal (mean± standard error).

Table 3.2: Comparison of initial andfinalcondition indices by feeding schedule 34 and stoCking density treatment (mean±standard error).

Table 3.3: Morphometries of 0+ wolffish, held at 9°C. fed six dietswitha 43 range ofPE:'ffi values, throughoutthe12 weektrial(mean± standarderror). Table A:Length(cm); Table B: Weight (g).

Table 3.4: Morphometries of0+ wolffish held at ambient tcmpenture, fed six oW dietswitha range ofPE:TE values. throughout the 12weekmal (mean±standard error). Table A:Length(em); Table B: Weight (g).

Table 3.5: Totaltankweights of 1+ wolffish held at ambient temperatun:. fed 46 six dietswitharangeofPE:TE values, throughout the 12weektrial (mean±standard error).

Table 3.6: lnitial andfinaJcondition indices of wolflish fed six dietswitha 4748 range o£PE=TE values for twelve weela (meantstandardcnor).

Table A: 0+fish held at 9°C; Table B: 0+ fish held at ambient tcmpenuure;Table C: 1+fish held at ambient temperawre.

Table 3.7: Equations. R1 values and p values for gastric evacuation rates of 7J juvenile wolffish fed six dietswitha range of PE:TE values at 9°C. A: Stomach contentsingrams;B: StomlU:h contents as %BW.

Table 3.8: Equations. R1 values and p values for gastric evacuation rates of 76 juvenile wolffishfeda commercial salmon feed at three temperatureS.

A: Stomach contents ingr3mS;B: Stomach COntents as %BW.

viii

List of Figures

Page Figure 2.1: Schematic diagram of the recirculation facility. A: header tank; 14

B:rearingtank; C: filter (fibreglass bedding);0:filter (hio-rings);

E: sump tank; F: pump; G: chiller unit; H: micro filters.

Figure2.2: Experimentaltankarrangement in the recin::u1ated seawater facility. 16 Figure 3.1: Mean specificgrowthrates (SGR) of0+ wolffish held at three 35

stocking densities and fed according tothreefeeding schedules at 9°C.Verticalbarsrepresent standard error.

Figure3.2: Mean meal size (mg feedfg fish) of0+ wolffish stocked at 35 three densities and fed according to three feeding schedules at9°C.

Verticalbarsrepresent stanclard error.

Figure 3.3: Mean feed conversion ratios (FCR) of tanks of 0+ wolffish held at 37 three stocking densities.Q(20gIL)=21; n (50gil and 80gil)=26;

Verticalbarsrepresent standard error. (. denotes statisticaUy different values).

Figure 3.4: Mean protein efficiency ratios (PER) oftanks of0+ wolffish held 37 at three stocking densities.n(20gIL)

=

21;n(50gil and80gil)

'" 26;Verticalbars representstandarderror. (. denotes statistically different values).

Figure 3.5: Temperature profile of recirculated, thermostatically controlled 39 system (St. John's) and the ambient seawater at Wesleyville, Newfoundland. Startingdate: September 10, 1995.

Figure 3.6: Weight change ofwolffish fed six dietswitha range ofPE:TE 41-42 values under two temperature regimes. [A:0+ wolffish held at 9°C;B: 0+ wolffish held at ambient seawater temperaeure.] Mean weightingrams.

Figure 3.7: Weight change of 1+ wo1ffish fed six dietswitha range ofPE:TE 45 values at ambient temperatures. Total weightper tankin grams.

(pE:TE "" 0.35, 0.45 and 0.46 eachhadone mortality between week S andweek.8. PE:TE=0.35hadtwo morta1ities between week 8 andweek12).

Figure 3.8: Specific growth rates (SGR) of0+and1+wolffish fed sixdietswith 50·51 a range ofPE:TE values UDdertwotemperatureregimes.fA:0+

wolffish heldat9"C;8:0+wolffisb heldatambient seawater temp- erature.;C: 1+fish atambienttemperature.}Verticalbarsrepresent standard error.

Figure 3.9: Mean dailyfeed intake (rog feedlgfish)by 0+and1+wolffish fed SJ.S4 six dietswitha range ofPE:TEvaluesundertwotemperature regimes. (A:0+wolffish heldat9°C; B: 0+wolffish held at ambient seawater temperatUre.;C: I+fish at ambient temperature.]

Verticalbarsrepresentstandarderror. SiInilar letters denole statistically similar values.

Figure 3.10:Mean daily lipid intake (rog Iipidlg fish) by0+and1+wolffish fed 56-57 six dietswitha range ofPE:TE values undertwotemperature regimes. [A:0+wolffish heldat9 DC; B:0+wolffisb held at ambient seawater temperature;C:1+ fishatambient temperature.]

Verticalbarsrepresent standard error.Similarletters denote statistically similar values.

Figure 3.11:Mean daily energy intake (calorieslg fish) by0+and 1+wolffish 59-60 fed sixdietswith a range ofPE:TE values undertwo temperature regimes.fA:0+wolffish held 819 DC; B:0+ woLffishheldat ambient seawater temperature; C: I+fishat ambient1emperalUn:.J VerticalbarsreprescntstaDdarderror.

Figure 3.12:Hepatosomatic indices(%BW)of0+wolffish, heldat9"C.in 62-63 relationtothe dietaryPE:TE value. Verticalbarsrqnesent standarderror.

FiCUre 3.13: Hepatosomatic indices(%BW)of0+wolffish, heldat9°C.in 62-63 relationtothemeandaily lipid consumption(mglipidlgfish).

Verticalbarsrepn:sent standard error.

Figure3.14:Hepatosomatic indices(%BW)of0+wolffish, held at9DC.in 62-63 relation to the mean daily lipid consumption (rog Iipidlg fisb).

Verticalbarsrep~seDtstandarderror.

Figure 3.15:Fet:d conversion ratios of0+and1+wolffishfedsix dietswitha 65-66 range ofPE:TE values undertwotemperature regimes. fA :0+

wolffish held at9DC; B:0+wolffish held at ambient seawater temperatUre;C:1+fish at ambienttempen.ture.]Verticalbars representstandarderror.

Figure 3.16: Protein efficieoc:y nllios of0+ andt+wolffisb fed six dietswitha 68-69 rangeofPE:TEvaJuesunder two temperatUrerqimes.[A: 0+

wolffi$h held at 9"C;B: 0+ wolffish held at ambient seawater tem~;C:1+fish atambient temperature.] Verticalbals repleScntstandanierror.

Figure 3.17:Costofproduction($Ikg) of0+ and 1+ wolffish fed six dietswith 70-71 a taIlgeorPE:TEvaJuesunder two temperuureregimes.[A:0+

wolffisb held at 9°C; B: 0+ wolffish held at ambient seawater temperature; C: 1+ fish at ambient temperarure.) Vertical bals representstandard error.

Figure 3.18:Comparison ofgasuic evacuation rates (gIhr) of0+ wolffish at 74 9°Cinrelation to dietary PE:TE value. Initial meal size fixed atO.10g.

Figure 3.19:Compari$Oo of gastric evacuation rates (%BWIhr) of0+ wolffish 74 at 9 °Cinrelationto dietalY PE:TE value. Initialmealsizefixed atI.OO%BW.

Figure 3.20: Comparison orgasttic evacuationrates(gfhr) of0+ wolffish at 77 thR:e tempenltures.Initialmeal sizefixedat 0.25 g.

Figure 3.21: Comparisoo ofgastric evacuation rates (%BWIhr) oro+ wolffish 77 atthreetemperatures.Initialmealsizefixedat 2.00 %BW.

List of Appendices

Page Appendix A: Weightsandlengths ofjuvenile wolffisb by year class and 103

rearingmetbod.

Appendix B: Thtcc way ANOVA for the SOR of 0+ wolffish fed commercial l04 salmon feed bystockingdensity, feeding scheduleand~licates at9°C.

Appendix C: Four way ANOVA for the daily feed intake of0+wolffisb fed lOS commercial salmon feed by stocking density, feeding schedules, periodand replicates at 9

0c.

Appendix0:Thtcc way ANOVA for the FCRofO+ wolffish fed commercial 106 salmon feed by stocking density, feeding schedules and replicate at9°C.

Appndix E: Three way ANOVA forthePER of0+ wolffish fed commercial 101 salmon feedby stocking density, feeding schedules and replicate at9 OC.

AppeadixF:Three way ANOVA forthelength of 0+wo1ffishfed six for- 108 mulated diets with a range ofPE:TEvalues.over thnc·four weekperiodsandtank replicate at 9"C.

AppeadixG: Thtcc way ANOVA for tbe weight of 0+ wolffish fed six for- 109 mulated diets with a range ofPE:TEvalues,over three-four week periods andtank replicate at 9"C.

AppeadixH: Two wayANOVAfortbcarcsine ofthe SGRofO+wolffisb 110 fed six formulated diets with arangeofPE:TE values over th:ree-four week periods at ambient temperature.

Appendix I: Threeway ANOVA for the feed intake by 0+ wolffisb fed six III formulated diets with a range ofPE:TE values overthreefour~

week periods and replicates at ambient temperature.

AppendixJ: Two way ANOVA for the feed intake by 1+ wolffish fed six IU formulated diets with a range ofPE:TE values overthree periods at ambient temperature.

xii

1.0 Introduction

Thestriped wolffish is one of a number of cold-adapted species being investigated for its aquaculture potential in thecoastalareas of the North Atlantic,. including Newfoundland.

Wolffish belong to the family Anarhichadidae and arenatiwto the North Atlantic and PacificOceans(HubbsandBarnhart.1944; Wilimovslcy, 1964). The three species that inhabit Newfoundland waters are the striped or Atlantic wolffish,Anarhic/uu lupusL.

the spotted wolffisb,A. minor Olafsen. and the northern wolffish, A. dentiCldatus Kroyer (Albikovskaya, 1983). Allthreeare common bycatchesinottcr trawls and gillnclS, but the northern wolffish, also known as the jelly cat, is discarded due to poor flesh quality (Tcmplcman, 1966; 1984).1be spotted wolfish is the least common of the three. It rarely inhabits water less than 50 m deep or warmer than 5°c(Albikovskaya, 1982). The frequency of capture of northern wolffish in trawls increases as depthincreases,between 151 and 600rn,withsomereportsof capture in water up to 750 m deep. Northern wolffisb are most frequently captum1inwater colder than 5 'C (Albikonkaya, 1982).

Anarhichas lupus,alsoknownas the commonwolflish.is most commonly found in water up to 350 mdeep and colder than 4

'c

(Albikovslcaya, 1982).

f. fBiological Adapttltions

Wolffish display a range of adaptations to a cold., inshore habitat which are of particular interest to aquaculturists.Theseinclude tolerance to cold temperatures by employment of an anti-freeze protein, large cggs, advanced development of larvae at hatching and a relatively minor mctamorphosis.

1.1.1 Temperature Tolerance

Tolerance to low water temperatures is of concern to aquaculturists, especially those in northern locales. Ambient seawater temperatures around Newfoundland regularly fall

below 0 °c during the winter. Those farming a species which cannot tolerate such temperatures are obligedtoheat the water, adding considerable expense to their operation.Anoption is to identify a species which not only survives low temperatures, but maintains a reasonable growth rate. One such species is the wolffish.

l.J.l,} Eggs and Larvae

Temperarure tolerance has been investigated by a number of workers. Pavlov and Moksness (l994b) compared striped wolffish egg development at six temperatures ranging from 5 °c to IS °C. They showed a decreased proportion of normally cleaved eggs at temperatures of 11°C and higher, A subsequent study examining four temperatures ranging from 9.9 °c to 15.7 °c led to the conclusion that 12.8 °c is likely the upper temperature limit for the incubation of wolffisb eggs (pavlov and Moksness, 1994b). Incubation of fertilized eggs at 1.0, 3.0 and 4.8 °C demonstrated that the lower temperature limit for proper development is likely 3.0 °c (pavlov and Moksness, 1994b).

The effects of temperature on hatching success were studied by maintainingtwogroups of eggs at a constant temperature of I °C or 3 °c and another at 6.5

0c.

Approximately two weeks before hatching the eggs held at 6.5 °c were changed to 7 °c or 10°C water.

Those eggs held at 7 °c and 10°C could not hatch without mechanical pressure on the eggs, so 7 °cwasdetermined tobethe upper limit for successfuJ hatching (pavlov and Moksness, 1994b). The hatchability of eggs incubated at I °c and 3 °c was 98.0% and 91.00/0, respectively. Thisisnot unlike Atlantic salmon which have a lower incubation limit between 0 °c aDd 1 °c (Wallace and Heggberget, 1988),

J.l.l.1Juvenilesand Adlllls

Few temperature tolerance studies have been conductedwithjuveniles and adults.

Moksness (1994) collectedwoLffishfry in the Barents Sea and maintained them intanks at ambient temperature for a minimum of 54 mooths. Based on temperature records and

data from monthly weigbings. he concluded thatlheoptimwn tmlpcranue forwolffish culture isbetween1"Cand 9 °C.Inthe wild. wolffish live at lemperatures between -1.9

°c

and 9

°c.

but geoera.Ily inhabitwatercooler than 4

°c

(Albikovskaya, 1982). This temperature tolerancerange differs from Atlantic salmon which tolerate temperatw'eS between5.5

°c

and24

°c

but grow most effectively between 10"Cand11

°c

(piper~(al 1986). The optimal temperature range for wolffisb falls within normal ambient lemp«atw"eS for the Northwest Atlantic Ocean.

1.1.3 Fecundity

Anunderstanding of the physiology of reproduction in wolffisb is important for establishing and maintaining productive broodstock. Wolffisb fertilization is internal, so themotility of undiluted sperm can last for severaldays(at 0104 °C) while the known duration of viability is approximately 10 bolUS (pavlov, 1994; Moicsness and Pavlov.

1996). Salmon spenns.reviable for a minute or less (pavlov and Moksoess, 1994b).

Wolffish produce a small amount of sperm. wbicb is less dense thanthat of Atlantic salmon (Kazakovand Obtazsov, 1990; Pavlov andRadzilcbovskaya,1991). Two main tl:producrive strategies are evidentinoatUre.1be r-strategy, evident in pelagopbilous fish suchas cod, involves producing Large numbers of youngaDd providing little or no parental care. Thek-stn1egy, common to Atlantic salmon(Salmo saJar),Iwnp6sh (CyclopUT"US lumpus)andwolffish,involves the production ofsmallernumber of young.

witb.agreaterdegreeofparentalinvestment (SaiP~toJ.,1986).Thisinvestment, in the case offisb.,may comeas a larxe egg with ample yolk reserves., or can:in!he fonn of guarding of the nest and free-swimming young. This isinmarked contrast to the strategy of producing hundreds of thousands to millions of pelagic eggs and releasing them for distribution by ocean currents. Turbot(&ophthalmus maximus), for example, produce 1 million eggs per kilogram annually while wolffish and Atlantic salmon produce roughly 2000 eggs per kilogram annually (Tilseth, 1990; Pavlov and Momess, 1994b). Male wolffish guard the egg masses and ensure sufficient water circulation through the mass

(Keatset al., 1985; Ring0 and Lorentsen, 1987). Pavlov and Moksness (1995) detennined the incubation period (fenilization to 50% hatch)indays(y)to be approximated by the equation:

y

=

425.28 - 77.875x+6.584x²-0.20325Xl, (1)

where x is the temperature('C).Wolffisb larvae hatch at an advanced stage of development, ready for exogenous feeding and closely resemble small adults.

Establislunent of husbandry conditions for successful reproduction of captive broodsiock will eliminate the need for yearly collection of wolffish egg masses.

1.1.4Egg Size and Larval Developmenl

For aquaculture purposes, eggs shouldbe large and produce well developed larvae (Soin et aI., 1986; Tilseth et aI., 1992). The size of wolffisb eggs facilitates management and naodling associated with incubationandwolffish larvae hatch at a more advanced stage of development than do most marine fish (Table 1.1). Halibut larvae(Hippaglossus hippoglossus), for example, are still embryos when they hatch, having no functional eyes, jaws or gut (Pittmaner al., 1990). Wolffish begin exogenous feeding on Artemia within hours of hatching, although stan·feeding within six days of hatcb is aonna! (pavlov and Novikov, 1986). Wolffisb readilyweanon to anificial feeds (Moksnesset a/., 1989).In fact, Strandet aI. (1995) start·fed larvae using a floating, fonnulated feed in a shallow raceway and bad 82% survival to 60 days post-hatch. Yellowtail flounder larvae (Limanda ferrug;nea) begin exogenous feeding at 4--5 days post-halch andrequire copepod nauplii <100jJJIlinsize (Smigielski, 1979). They graduallyweanonto larger organisms (adult copepods. rotifers andAnemia sp.) as theygrowand metamorphose (Smigielski. 1979). The metamorphosis between the larval and juvenile stage is very minor in wolffish. Known as direct ontogeny, this characteristic is desirahle in potential

Table 1.1: ComparisoD of egg diameter (rom) and body lengths (mm) at hatch for several cold-water marine fish species.

s... [uou..-eter LeaJlllat Haldl

AlilalltieCod 4.4:t0.1' GoteeiwandBrown, 1993

(~",orlrwtJJ 1.2-1.6 3.5-4.0 TIIJetb.I990

1.1-(.4 1.' Ro$en[lIlldaal~199]

Atla.nticSalmo.¹ 5.0·6.0 15.0-20.0 Tilseth.l990

(SaIMOsaJ",) 4.6-U Thorpural.,19S4

Halibut(HippoglossllS hippoglOSSlU) 3.0·3.5 6.5·7,0

.. ^,

Tilseth.l990^{Pittmant'laI.,1990}

Lumpr15b 2.6:tO.3 5.1;1:0.3 BrownetaI.,19'n

fCycloptuwlUlltplU)

OctaD Poul

(MacTO%l1Q1"CGamuieamIs) 1.5tO.2 39.2:1:0.7 Brownetal~19'n

TllrtlolfScoplrdJahr.llS _ _a) 0.9-1.2 2.7-3.0 Tilseth.I990

WoItre,. 4.7·5.1

"

PavlovCldMotslless..I994a,b

(A_hidtaJ/upKI) 4.7-6.0 20 PavIovCldNovikov,I916

5.5·6.0 20 Titsem, 1990

Ydlo..-uUflolIa4tr 2.75 SmiJielski,I979

(Limonda/nnigiNDJ 2.66 LaurenceandHowell,1911

0.79-1.01 Cohoo and Matak,1969

0.61·0.76 2.[-7.0 Tilseth.l990

ISto7dayspoSl_lwcb

1 AtlaoticsalmonilRillWlrolllOlU,spen<tillgtheiradult live$in$QWI%et",eltc:eIX for spawninl llIigflnions inlOfreshwaterstreanllandrivm.

aquaculture species since metamorphosis may cause many mortalities(Balon. 1985; Sain ttaJ.•1986; TiIseth. 1990).

Repons ofcannibalism.inwolffishexist (MoJcsness, 1990), but are not commonplace (Pavlov and Novikov, 1986;Ring+ttaI.,1987; Mokmess ttaJ.,1989). Cannibalism appearstobe linked to low stocking deosities and to a large size gradient among fish in a lank(Moksness and Pavlov. 1996).

U.SGrowth

Juvenile wolffisb have been raised in Norway at an ambient temperature ranging from 6.6 •to11.7°c. They werehand-fed moist ordrypellets, with a proteinenergy:total energy (PE:TE) between 0.50 and 0.63 (StefanusscnItaI.•1993). Those fish feddry pellets bad a significantly highergrowthratethanthosefed moist pellets (026 i/dayand 0.15 i/day. respectively: Stefanussenttai.•1993). Wolffish larvae have exhibited maximum specificgrowthrates (SGR) up to 3.34% OW/day dwing thefust 108 days post-balchand up to 3.6% BW/day from 100 to ISO days post-hatchatambient temperaD.1rCS(MoksnesstlaJ.,1989).Thesegrowthrateswereachieved using threedry diets and different feeding regimes and weaning schedules. At optimum rearing conditioos. (below 8"C) MoksnessIIaJ.(1989)claimedthatstripedwolffishcould reach 2.5 kg within 2 years ofhatcb.ing.

1.2 Culture"'_tho,"

Thestripedwolffish, with its lean, fine-textured flesh, toleranceto low water temperatures and its hardy, well-developed larvae. is a prime candidate for culture in Newfoundland (SomItoJ., 1986; Seafood Leader, 1991). The challenge lies in determining the husbandry practices which willtnaximiz.ethe health,growthand reproductive capabilities of the fish and. ultimately, the profitability of the aquaculture

venture. Tbe egg rearing protocol and the care of larvae are well established and researchers in Newfoundland haveachieved 93.5% survival of larvae through mewnorpbosis using high light intensities., altered photoperiods and artificial feeds (Wiseman and Brown, 1996). The next step is determining the protocol for rearing ju\ocniles.

1.2.1 Stocking Density

Wolffish live solitary lives. normally inhabiting rocky crevices. except when they pair up during spawning season (Moksness and Pavlov, 1996). This raises questions about the potential co culnue this species at economically viable stocking densities. Some species do not adapt well Cohigh stocking densities. Veryhighstocking densities can lead to aggression, increased risk of disease and poorgrowthrates (Keenleyside and Yamamoto.

1%2; Fenderson and Carpenter. 1971;Baker and Ayles. 1990). However, Arctic chan' rearedat high stocking densities have higher mean weights and lengths than those reared at lowerstOCking densities UDder the same cooditions (BrownItaJ.,1992). This may result from the reduction of antagonistic interactionsbetwoeeDfishandthedecreased time spent5'Nimming. thus decreasing the expenditure ofenergy.

1.2.2Diet

[nits natural habitat,thesuipcdwolffish consumes a variety of prey similar to those consumed by spotted wolffish(A. minor) (Albikovskaya. 1983).TheseincludeLithodts spp. and Hyas spp. (Oecapoda). sea stars (Asteroidea). brittle stars(Ophiura).

Srrongylocentrorwspp. and sand dollars (Echinoidea). clams (Bivalvia) and whelks (Gastropoda) (Albikov$kaya, (98]; Templeman 1986). Indigestible body armour protects many of these prey itelIlJ. Wolffish teeth cannot grind down this skeletal matter, so they crush it to facilitate swallowing. OrlovaItaJ.(1989) provided wolffishwithmeals of natunt! prey items, somewiththe shells removed. From their observations of gut passage

times theyconcludedthat.geoeraJ.ly, the more indigestible matter amc:aJ.contaioed.the fasteritpassed throughd1estomachand intestine. However, whenafood boluspassed quickJy through theintestinethedigestible components oftenpassedundigested.

Feedingnaturalfoodisnot an optioninthecontext of intensive culture, therefore, itis necessaryto developandevaluate formulated feeds. Whiletherehavebeena number of studies of larval wolffish nutrition. therehave:been few investigations of the dietary requirements of juvenile and adult wolffish(Rin89et al., 1987; Moksness et al.• 1989;

Odovaet al.• 1989 and Moksness, 1990). These studies generally used natural prey, such asArtemia salina, scallops, mussels,squid,shrimp and echinodenns or commercial salmon feed. Some researchers have formulated their own feeds (Moksness. 1990), but the precise dietary requirements of wolffishareDOtyet known. Studies of other marine carnivorous fishhave:shown requirements for minimum dietary levels of50--60".4protein.

IG-2Q'l/o lipid, andmaximumlevels ofcarbohydrates, ash and fiber 1()...20%, 4%.andI().

250/.. respectively onadrymatterbasis (fucker. 1992; WlIson, 1994).

1.2.J Protein RequinlDenti

Carnivorousfishspecieshave:highrequirements for dietaryprotein(fucker. 1992).

Atlantic salmon(SaI1Msa/ar), for example.require45% of the dry dietas protein(Lall andBishop. 1977). Plaice(PleUTon.ectupia/usa)fedsix diets wilb. a range ofprotein from20to70%(drybasis)demonstrated. optimum weightgainwhenfed a diet containing 50%protein(Cowcyet al., 1972). Animalproteins offer the most nutritionally complete spectrum of amino acids sotheyare the most commonlyused protein. However,theyare very expensiveand drive up the cost of feeds, so measures 10 decrease feed. costs areneeded.First, since excess protein is metabolized for energy or deaminated and excreted, it is wise to determine the protein requirements of the species in question to avoid wastage due to inefficient usc. Adronel aI., (1976) demonstrated with turbot(Scophrhaimus maximusL.)that at comparable energy levels, a diet with a lower

8

protein content is superior to a dietwir.h a higherproteincontent. Therefore, it appean;

that the crucialfactoristhemtio of protein energy to the totalenergy(PE:TE).

Carnivores digest carbohydrates inefficiently. The requiredenzymes for carbohydrate metabolisman:present,butinverylow quantities. compared 10 Ilcrnivorous fish. This is evident in cod which. at low dietary carbohydrate levels have a relatively high carbohydrate digestibility. A5 the dietary level of carbohydrates increases. the digestibilitydropsdramalicany (Hemre~laJ.•1989). Therefore, with wolffish.th~

concern is 10findan optimum balance between protein and lipid while keeping carbohydrate values 10 a minimum.

1.2.4 PE:TE ReqttiremeatJ

Jobling~taJ.,(1991) found that adultcod gain weight rapidly when fed dietswithPE:TE - 0.40 10 0.45, without deleterious effects on the htpatOSomatic index. Younger cod (50 to 300 grams) show the best results when fed dietswithPE:TE~0.56 (Lie~laJ., 1988).

Young plaice(PJ~UTOMclupJaussa)fed a series of formulated feedswitha PE:1C range from0.14 to 0.92 showed optimumgrowthwbeofeda dietwithPE:TE-=0.70 (Cowey et 01.•1972).Ina similar study using turbot(ScopJuhalmvs maximvs L.)anda PE:TE range of 0.50 to 0.85 (isoproteic diets with decreasing IOtal energy contents), the authors concludedthatthe protein efficiency ratio (PER) reached an optimum valuewhenthe PE:TEwaslesstha.a0.50 (Adron~lai.,1976). An investigation of the effects of two waler temperatures co the protein requirements of juvenile sea bass(Dicelll7'archus Jabrax)wasperformed.. Four dry dietswithPE:TE - 0.162, 0.214, 0.263 and 0.313 were fonnulated andfedas a fixed ration tofishheld in a recirculated system. Atboth temperatures, IS and 20 DC, the highest weight gainwasobserved when PE:TE .. 0.263.

However, based on nutrient and energy utilization efficiencies, a PE:TE=0.214 was recommended for culturing juvenile seabassatbothIS

°c

and 20 DC (Hidalgo and Alliot, 1988).

1.1.4./ HepatosolfUl1ic Itldu

Thebepatosomatic index (HSl) provides information about the appropriateness of the energy content of the dietwhichcomplementsgrowthandpc:rfonna.ncedata. Atlantic salmon. for example, are able to store excess lipidsinvarious locations in the body, such as flesh and liver. They are abletoquickly mobilize these lipidR::SerVe$when needed.In non-oily fish suchaswolffish and cod, excessenergyis deposited as fatinthe liver. This may impair Liver function, wastedietaryresoun:esand reduce fillet yields.Anexperiment wilhjuvenile cod(Gadus morhua)wascompleted usingsixfonnulated diets withPE:TE from 0.11 to 0.61. A positive linear relationshipbetweenHSI and dietary lipidCOAten!

wasdemonstrated (lieetat.,1988).Infeeding trialswithturbot(Scopthalmus maximus L) using seven dietswithPE:TEbetween0.50 and 0.85. the lipid content of thefish

inc~asedas thePE:TEdecreased. However,atDOtimewasthecarcass lipid content of experimental fish higher than that nfwildfish(Adronet al., 1976).Thisresult isin sharp contrast totheresultsof an experiment with juvenile plaice(Pleuronectes platusa).

Using a dietaryPE:TErange between0.14and0.92overa12 weekperiod.fishfed each diet showeda significantly higher carcass lipid contentthanwildplaice(CoweyetaI.•

19n).Theseworkersconcluded that while the carcass lipid levelswererelated to the diet lipid levels, the total dietary energy content liIcely exceeded requirements.

l.loS FccdiaC FrtqllebC)'

Thefrequency of feeding depends, primarily, on the rate of gastric evacuation. However, a mUltitude of factors influence the rate of gastric evacuation, including energy content, panicle size, digestibility of the food, structure of the alimentary canal, water temperature and the phase of digestion (Tyler, 1970; Groveet ai, 1978; Persson, 1982; Bromley, 1987; Bromley, 1988; Dos Santos and lobling, 1988 and lensen and Berg, 1993).

Determination of an appropriate feeding fRquency is critical in order to minimize feed wastage andmaximizegrowth..Feeding too frequentlyresuJtsin wastedfeed, even if

'0

feedingbyband. Conversely, if meals ace too infrequent. the rate ofgrowthcouldbe compromised. Orlonet aJ. (1989), studying striped wolffish, documented food passage:

timesthrough the: gut rangingbetween4 and 10days for natural. foods such as scallops with and without shells,mussels, fish and shrimp. Highly digestiblefoodite:ms like:

formulated pellets could, theoretically, taketodays to pass through lhe gut. Halibut, (Hippoglossus hippoglossus), voluntarily eat large meals two tothreetimesperwee:k while the lemon sole(Microstomus/cit(Walbaum) consume small meals every one to one and a half days (Dave:npon et al., 1990).Theauthors attribute this to the fact that halibut posseslarge stomachsand short intestines, as opposed tothe:small stomach and looping intestine of the lemon sole. WoLffish have: small stomachsand intestineswhichme:asure ana~of68% oflbc:bodylength (Verigina. 1974). This definesthe:Deed to establish Ihe rate of gastric evacuation and return ofappetite.

Thereareseveralways to evaluate lhe rate of gaslric emptying infish. One common rnelhod is to feed fisha meal containing contrast media such as barium sulphate. The progress of the meal (often pre-weighed)through the gastrointestinal tract is rc:<:orded by periodicx-rays(Edwards,1971; Joblinget aJ., 1977;Groveet aJ., 1978; Flowerdew and Grove, 1979). This technique isaccurateanddoesDOtrequirekilling large numbers of fish. Anothercommonmethodis to seriallykillaspecifiednumberof fish atpre- determinedinterVals following a meal and remove, dty andweighthestomach contents Uobling, 1980;MacDonalderai.,1982;Persson.1982;Bromley,1987; Jensen andBerg, 1993). A third method isthe:anaiysisofstomaclt contents pumped out ofananac:sthetizc:d fish ata given timeafterapre-weigbc:dmeal (Bromley, 1988; Dos Santosand Iobling, 1988).

1.3Ratlona'.

Sincethe moratoriumODthegroundfisbetywasestablished inJuly1992,fresh,white- fleshedfishare inhigh. demaDd and are sold at a premium price.The:price offreshcod

11

(Gadus morhua) in Newfoundland markets now exceedsthatof salmon,whichwasonce considered a delicacy. Thisbasencouraged aquaculturists andresearcher.; to identify a suitable species foraquaculturethatwill providea sourceof fresh, wbite*fleshedfISh.to addition tohighquality flesh, this species mustpossesscbaracteristicsthat faciliwe culture. Many marinespecies have a precariouslarval developmentaMexhibitmass monality during metamorphosis. This iscurrentlyanareaof extensive researcb. An alternative to resolving these issues is to identify a specieswithlarge eggs andwell- developedlarvaeat hatch. Thesean:hforsuch aspecieshasresultedin special attention being given to the wolffishes. The primary marine fish species cultured in Newfoundland currently is the Atlantic salmon. Except for a series of farms using heated effiuent from a hydroelectric plant. low temperatureshave resulted in slowergrowthrates thanin other, more southerly, locations. Atlantic salmon also have a weU.understood, relatively simple larval stage. Newfoundlandbasanobviousneedandcapacity for a water-basedindUStry.

Itappears asthough aquaculture is the idealindustry to fill that niche. With ambient temperatures whichareconsidered lowby thestandardsofproducersin temperate climates,the culture of ahardy, cold-water species. such as wolffish must be developed..

Withthecxt:raordirwy success oflarvalwolffisbculture to date,thereis aneed for development of a protocols for the culture ofjuvenilesandadults.

The purpose of my studywastoexaminethefoodand. feeding requirements of juvenile stripedwolffishunderculture conditiooswiththeaimof maximizinggrowth. The objectiveswerethreefold:

I.Establish a suitable stodtiDI density for juvenile striped wolft""ub.

2. DetermineaDappropriate fHdiDe schedule for juveDile striped wolfrub.

3. IdentifyaDeffedive nnee of dietary enel'l)' balances for striped wolffisb diets.

12

2.0Materials and Methods

2.1 AquariumFltCiHtiu

Experimentswereconducted intwofacilities. Thefirstwas thelaboratory at the Fisheries and Marine lnstitule of Memorial UnivetSily of Newfoundland in 51. lohn's.

which has a seawater recirculationsystemwithtemperafUreregulation.Theother facility usedwasthe Wesleyville Marine Fish Hatchery in Badger'sQuaY.Newfoundland, which is a flow-through system operating at ambient sea water temperature.

2.1.1 Recirculation Facility

A recirculated seawater facility was used for the research completed in St. John's. This laboratory was selected because of the temperature control system whichmaintainsthe water temperature within0.3 OCof the set temperature. The labwasdivided intotwo sides. each of wb.icb contaioedsix -750 Lgreenfiberglasstanks(Figure 2.1). These tanks""'ereset up ontwOlevels.,with threetubsoneachlevel. Tankshad individual water supplies and individualdrains.Thedrainsjoined into a common pipe on each level and the levels then combined to return to the sump through a series oftwopaired biological filters. Two of the fourfilters werereplaced every twowedtssothat the bacteria population on thefilter was maintained.Thetotal capacity of thesystemwas2.7 m^J. Waterwastransported.by truck:fromtheOceanSciencesCentre, situated00the shores of Logy Bay. Weekly replacement ofwaler in the system was approximately 2o-li.

Replacementwasgradua1,with seveta1gallons being replaced daily. Oncea month a flushingoccurredwhen one third of the water in the systemwasreplaced at once. The tank.floors.below the suspended experimental containers. were siphoned regularly to optimize water quality.The salinity wasconstant at32ppt. The dissolved oxygen concentrationwasnever lowerthan92%in thetanks andat times the concentration reached105%sanuation, thereforeaerationwasunnecessary. Thepbotoperiodwas

13

Figure2.1: Schematic diagram of the recircu.lation facility. A. header tank; B. rearing tank;C. filter (fiberglass bedding); D. filter (bio-rings); E. sumptank; F.

pump; G. chiller unit; H. micro filters.

'4

maintained at18l:6D. The light intensity overthecourse: of

me

experimenlwasCOOSWlt wilh a mean of 116.51ux (S.E. - 14.5).

Experimentsinthe~irculationsySlem were conducted eitherintanks or baskets which were modified to accommodate fish. The tanks used were white4.5 L plastic containers wilhanoutflow hole cut4 cmfromthe bottom. fiberglass meshwasglued over the outflow50fish and feed would001escape.Sixoftbese containers were strapped together in a circle using electrical ries andaseventh, bottomlesslankwasattached inthe centre10 act as ballast, keepingthetankslevel at all times(Figure 2.2). The inflow pipewasfined with six spigots which,~coonec;:lCd to pieces of plastic tubing, delivered water to each t.ank.. The water temperaturewasmaintained at9.0 :t OJ °C throughoutthefeeding trialsbyuse of a thermostatically regulated chiller unit (Johnson Controls., Milwauk.ee, Wisconsin).

The baskets were pale bluerectangularplasric containers measuring23cm by35cm by 15 cm. The sides and bottoms of thesepanswere cut out leaving a frame on whichwas glued fiberglass mesh. The mesh (5t8ndardmosquito screeo)wassmall enough to prevent feedfrompassingthroughthe tankfloor. This is critical since wolffishare bottom feedenand thetola! amount of feed consumedwasrecorded.at each meal.The basketsweresuspeodedfromwoodendowels such that theyweresubmerged towithin 2.5cm of thelaps. A mesh fencewas erected IIl'OUDdeach basket to preventfishfrom jumping out ofthebasIc:ets.. Stringwasthreadedthroughthe bottom oftbe feDCc: andtied around the top of thebaskeu. Fenceswereheld erect using plastic straws placedinboles in Ihe comers of the bask.ets andby tying the top edges of adjaceot fences together. Two baskets were suspendedineach tub, such that the volume ofwatercontainedineach basket was approximately 7L.The inflow pipe was fittedwithspigotsandtwo inflow hoses were directed into each basket, delivering atotalof 2 Umin.

15

Figure2.2: Layout of plastic tanks usedinthe stocking density/feeding schedule trial.

16

2.1.2 flow.TbnJugb Facility

Theflow.through facility at Wesleyville, Newfoundlandisan experimental marine fish halchery. It conta.i.m a larvalrearing roomwhich is abo suitable for juvenileculture.The system is made up of27greenfiberglass raceways of dimens\ons106emIf.2Sem x IS cm deep. The lenglh of the raceways and water depth were adjustable based on the requirements for each experiment. The raceways were built inlothret:racks. eachwith threelevels andthret:raceways per level. Waterwaspumpeddirectly from 8 m depth outSide the hatchery, throughasand filter to the larvalrearingroomwhereitwas distributedlOeachraceway.Theracewaysdrainedintoacommon pipeandwaterwas discarded.. Aplastic slatted screen separated thefishfromthe outflow, soneitherfood norfishentered. the drain. Therewas110watertempmuurecontrol in thelarvalrearing room. andthe ambient water lemperaturewasrecordeddaily. Dissolved oxygen levels were consistantly above90%saturation.Thelight intensitywasapproximatelytOOlux and the photoperiod was18L:6D.

2.2 Fish

Twoseparate year-classes offishwereusedintheseexperiments.Theolder (referred10 as1+for the purpose of this study). <:arne from twoeggmasses collectedbydivers off BaulineinConception Bay. NewfoUDdlandintheautumn of 1993.Theeggmasses were taken tothe recirculation syslem where theyweregently pulledapartandincubatedin4

°c

!!CawalerwhichwasbiologjcalJy filtered and exposed to ultraviolet light. The incubation unit (Heath tray)wascoveredinblack plastic tominimizethe exposure of the eggs tolight. Eggswerenot treatedwithan antibacterial agent, as recommended by Pavlov and Moksness (1993).Oneofthetwoeggmasses wasdestroyedbybacterial ir-.fection. Deadeggswerepicked daily. and when thefishbegan tohatchthey were placedinthe white tanksdescribedpreviously.lbese.the older ofthetwoyearclasses.

hada meanhatchdateof March 2.1994. Initially, larvae'NerefedAmmia enrichedwith

17

High DHA Super Selco (Arlemia SystemS.1NVE Aquaculture NV.Baasrode. Belgium) which isacommercial blendofpolyunsaturatedfattyacids. The larvaewere soon weaned onto aseriesof feeds designed for marine fish larvae and produced byLansy (INVE Aquaculture NV. Baasrode, Belgium). Feeding trials continued until the fish were 60days old. Following thelarvalfeedingtrials the fishweremaintained ontansydiets.

Atrial using moist feeds was then started. Having beenfedondryfeeds. the fish would

nOlaccept the moist feed. Despitethefactthatvarious binders and binding techniques wereused,thefeed consistently crumbledwhenchewedbythewolffishand,therefore, was not completelyingested.Forthisreason,allsubsequentexperimentswereconducted usingdry.extruded feeds. Whenoot beingusedinanexperiment,fishweremaintained on the Hi Pro SalmonGrowerDiet (Corey FeedMills,Fredericton, NB) (fable 2.1).This was supplemented withfresh,frozenfeed,suc:h as chopped herring orsquid.

When the experiments began with thet+fish,the group of fisbused inthe stocking density trial were approximately 530 days post-hatch. Those involvedinthe dietary energy trial were approximately 577 days post-batch.Themorphometric summaries of theyear classes are showninAppendix A.

The youngerfish(referred to as0+)were coUectedinNovember 1994andincubatedin thesame manner as!beolderfish. However,theseeggswere treated with g1utanlldehyde in ordertoprevent bacterial infection. Treatmentswereappliedwhentheeggsappeared unhealthy.Themean hatchdatewas February 2. 1995.Theywere:initially fed Anemia butw~weanedontoone ofthreecommercialmarinelarvae diets(Lansy;Biokyowa, Kyowa Hakko Kogyo Co.,lId, Tokyo, Japan; or Moore-Clark, St. Andrews, New Brunswick) WItH 70 days post-hatch.Aroundthat time the fish stopped swimming and eating. They were moved from the white buckets into the green tanks and were fed frozen. chopped herringand squid. Their activity level soon increasedand threeweeks later they wereweanedontothedryfeed onceagain.Fish hatched at the Wesleyville Hatcheryand the OceanSciences Centreexperieocedsimilar changesinbehaviourwhen

,.

TableZ.I:IngredientSandproximateanalysisof experimental diets.

Dlell

IDcreclicaul SalmDa

.Hd

Fish Oil

'''l

^22j

^I"

^{16.0 13.0}

...

CelhilO'le

''''

^16.7

... ...

^l.'

. ^.•

AnimillProtein Produets

'''l

^{37.0 42.0 47.0 S2.0 S7.0 62.0}

PllU'IlProtein Producu

''''

^{14.0 16.0 17.0 19.0 20.0 22.'}

J>roceuedGnin By-products (%)

...

'.1 '.1 '.2

'.7 7.'

Viwninsand MinetJls

'''l ^{I.' \.. \.. \.. I.'} \..

ProsimateAaa!ysis

Moimn (%,

'.04 .... '04

^{S.S3 W}

-~

^{(%, 36.04 40.62}

44'"

^{SHI SI.43 51.45}

Upid (%,

29."

^{27.7 26.71 23.67 21n 16.96 11.07}

CW>h""",' (%, 21.J7 1...2 lJ.OS 17..36 10.14 1l.14

"'"

^(%,

^.", ....

^{1.17 . j ' 1.93 10.91}

TotalEnerzy' (kaIIsl S.74 5.67 '.62 5.45

, ... _", '.06

PE:TE . j , '.40 ^0.45

....

^{O.SS 0.6\ 0.49}

I Oien 1-6were f<Wrllu'-cd andpreparedbylicgltt(OardnmPAl. The fonnullliOl'lJ ....trebasedOIIa,arnmm:iaJ salmon feed, lh=fal'C,dcl&ilcdingredienl information is lIl\&vailable.Thesalmonfeedill. propriewy 'ammm:iaJ di,t: HiProOfol«f Salmonid DiCl (COfC)' Feed Mills, Frtdc\ic;lon.N~Bnmswiek).

1 Allvaluesate given on adlydielbNis.

'Cart>ohydrateUlntallestirnatedbY$llbll'1dion.

·Total{g1nss)enel'CYvaJvesealQllMcdllSinlllanl1atdo;:aJoricval_(Cl>af!lo/..I9I2~Protein:S.6keal1a; lipid:

9.5 kcalla;c:ariIoh)'dnte:4.1kealia.Onekiloc:alorieisequalI04.1I4kilojoulc:s(Dorbnd, 1915).

"

theyrQCbed70 days post-hatch, andhighmortalitieswererecorded.Modifica.r:ion of the feed ortankswhen feed consumption patlem5 change mayboost feeding and activity rates and avoid mortalities.

Whenthe dietaryenergyTrialbegan, the 0+ fishwere209 days post-hatch. These juvenileswerefed more intensivelythanthosefishhatchedtheprevious year.and the mean weights and lengths reflect this (Appendix A). The mean weights differed by less than 2 gand therewasa fuji year between hatch dates.

2.3 ExperimentBl Design

1.3.1 Feeding ScheduleJScockiDg DenJIity Trial

Each. of three stocking del15ities. 20gIL.50gil. and 80 gil.,wasrandomly assigned to 9 round containers. for a total of 27 containers. Three containers of each group of nine wereassigned a feeding schedule; two meals/day, ooe meal/dayandonemea1Itwodays.

One blllldrNandfiftythreefish wererandomlyassigned to the27 containers to produce stoclcing densities as indicated. Fishwerehandfed according to the schedule assigned to their respective tanksandtheywereweighedandmeasured everyfour weeks foratotal of 12 weeks.

2.3.2 Dietary Energy Tn.1

Two hundred and forty - 0+ fisbwereevenly distributed among 12 basketsinthe recirculated system sothat the stocking densitywas20w'L. Thesix dietswereeach.

randomly assigned to two baskets. The feeding trial lasted 12 weeks. Fisb lengths and weights were measured at the start of the experiment and then at four week intervals.

Th.is Trialwasrepeatedin racewaysatthe Wesleyville Hatchery at ambient water temperatures. Athirdgroup of fish, 577 dayspost-hau:h.,wereassigned to 12 racewaysin

20

Wesleyville atthesamestOCking density. Thesix dietswererandomly assigned to raceways and IDOfllhologic:ai measurementswerelaken every four weeks.

The 240 fish usedinthe !rial at the recirculation facilityweresubsequently killed for stomach analysis dwing the gastric evacuation !rial. At that time. the liver.; were weigherl in order to determinethehepatosomatic index (liver weight as a percentage of whole bodyweighl).

2.3.3 GuttieEvacuation Trial

The serial slaughter methodwas chosen for the gastric evacuation trial because of the low cost and reported accuracy of the results. Following a period of starVation, fish were fed toapparent satiationand fivefish from each treatment wereslaUghtered initially and al8 hourintervals.Fish were frozenimmediately following slaughter.Thismethod pennitted removal and weighing of the livers to provide data on the changes in hepatosomatic indices due to variationsinthedietary energybalance.

2.).J.I TempowtllnSttul}1

Wolffish at the WesleyvilleHatcherylost their appetite when the ambient water lemperalUre reached 1

°c

(Watkins.pelS.comm.). Pavlov (1995) reported thatthe optimum temperature range formaximizing growthfatesinjuvenile wolffish is 10 DC to 14

Dc.

However, Moksncss (1994) reportedthatthe optimum rearing temperature for juvenilesis below 10 DC. Based on this infonnation, four temperatuIes were selected: 2 DC,S DC. 9 DC and 12 DC. These values encompass the most likely range of water temperatures encountered under Newfoundland's coastal culture conditions.

21

2.3.3.1 DwruyEn.ergy 8DJan.ceStudy

Six dim with a range of PE:TE values between 0.35aDd 0.61 (Table 2.1)wm:used to determinewhelherthe dietary energybalanceinfluenced the rate ofgastricevacuation.

The gross energy conlents of the dietswereapproximately equivalent, so this was ruled oulasa possible source of variationingastric emptying rate.Eachdiet was fed1040 previously starVed fish which wereanalysedusing the serial slaughter techniqueusing methodology describedinSection 2.6.2.

2.4 FeeM

Z.4.1DellJitylFeediDgSchedule Trial

Duringthistrial. the fishwerefed3.0mm.HiPro Salmon Grower peUets produced by Corey FeedMills(Fredericton, New Brunswick). The proximate composition is shownin Table 2.1. This is the same diet !bey had been fed forthepreviousthreemonths. Itwas selected because of its palatability.growthperformance and availability.

2.4.2 Die:tuy Energy Trial

For three IDOnths prior to thistrialthefish wereaUmaintained onHiProSalmon Grower pelletsproducedby Corey FeedMills.Preliminary observations of feeding responsesand grolNlh using pellets from threefeedmanufacturersshowedthat the Zeigler salmon stalter formulation appeared to be the IDOst palatableandproducedthebestgrowthrates (Zeigler. Gardners, PAl. Six diclSweredeveloped specificaUy forthistrialwith the assistance of Zeigler's Technical Services department. Zeiglerwasrequested to modify their salmon starter formulation to produce six diets with a range ofproteinenergy: total encrgy (PE:TE) from 0.40to 0.65. They modified the base diet by increasing the lipid and cellulose contentanddecreasing the protein content such that the gross energy remained relatively constant. Thecompositions ofthediets areshowninTable 2.1.

Zeigler cxpressed some concern over the pellet quality of the 10W-pIOiein diets, but all

22

diets had high quality pellets., aside from the high protein dietswhichproduced a considerable amount of fines. Thefinesweresieved off and lbc feed qualitywas otherwise excellent. The calculation of energy valueswasdone according tostandard valuesasreportedbyCoottaI. (1982).

2.5FeedingProtocol

Thereis some questionastowhetherobserverscandetermine when fisharcsatiated.

True satiation is the point when the stomachs of all fish under observation arefujI.Since this is impossible to determine without dissecting the fish, researchen must rely on observable behaviours to determine the point of apparent satiation (Coweyet ai.,1972).

Wolffish, for example, when they werehungrywould swim vertically with their heads protrudingabove the watersurfacewhen someone approached thetank,until foodwas offered or the researcher moved awayfrom Ihctank. When foodwasoffered, fIsh gradually sank orswamto the bottom, following feed pellets. As food continued tobe delivered., few fish remainedatthesurfaceandthe majority\o1r"C:fe00thebottom chewing pellets. Inearly stages of a meal, fishwhichwere inthe process of chewing and swallowing a pellet oriented thcmse:Ives toward anewpellet as itsankand even approached that pellet before thefirstpellet hadbeenswallowed. They frequentlyspatout theinitialpellet 10 take the new one if pelletswerepresented quickly.As the fishneared apparent satiation, approaching decreased first, then orienting decreased Atthatpoint.

fishDOlongerswamin thewater column.Thosefishwhichbad not yet reached apparent satiation generally rested on the bottom withtheanterior portion of their bodies propped up on their pectoralfins. Those whichweresatiated tended to rest completely on their ventral surface or lie on their sides.ThesepostUreSwerenot concrete indications of their level of satiation, but provided a clue as to how much more feed should be added. Based on preliminary observations,wolffisbdid DOt ingest pellets thatbadbeeninthe water more than one minute. Therefore, when fish no longer oriented, approached or ingested feed within a minute of presentation, theywereconsidered satiated.

23

Througbout these experiments6.sbwea: fed[Qapparentsatiation ratherthan being given a fixed percentage of body weighLPreliminary observations indicaled that some days very little feed was consumed. and other days an extr8Drdinary amount was consumed. If a fIXed percentage body weightwasdeliveredeachmeal.some days feed would be wastedand other days thefishwould be left hungry because theyrequireda larger meal than theprescribedamounL Inbothcases.the feed conversion ratio data would be distoned and thegrowthrates adversely affected. Afixed ration is not responsive[Q fluctuationsinappetite andgrowthis not maximized. Itwasforthisreason. too.that hand· feedingwas selected as opposed10automatic feeders. Hand feeding, though extremely time-consuming, allows theresearcherto monitor the feed consumption and identify changes which could indicate poor health. changes in water quality or feed requirements.

Experimental containen were siphoned prior to the morning meal, regardless of the feeding schedule assigned to a giventank.Feedingtimesfor fish fed twice per daywere9 a.m. and ) p.m.. Fish fed onceperdaywerefed at3 p.m.,aspreliminary investigation showedthatfeeding activitywashigherwhen the single daily mealwasprovidedinthe afternoonrather thanin themorning.Forthosefedtwiceperday, the morning mealwas the largest71.6% of the time (n"n9).Feed disheswere~ighcdinitially and following each meal. Fishwea:DOtfedtheday prior toeachweighingto ensure thatthe gutwas freeof significant amounts of feed and todisturbthe feeding regimeas little as possible.

Feeding resumedtheday following weighing. as scheduled.

2.6 Sampling

2.6.1 WeighiDg

Fish were weighed and measured one day prior to the start of me experiment and every 28 days thereafter over an 84 dayperiodfor atotal of four measwements. Fish were

2.

individually anaesthetised in a bath of25 mgIL MS-222 (Sigma Chemical Company, St.

louis,MO,USA). Eacb fisbwasremoved from the bath, and gently blotted with a paper towel, then placed on a scale with a ruler coveredwitha ttansparent plastic sheet. Mass (in grams) and standard length from the edge of the upper jaw when the mouth is closed, [0the tip of the nOlocbord(incentimeters)were lakenand the fish were briefly examined for signs of aggression (scars or torn fins) or disease before being placedinan aerated recovery bath.

Wolffish were remarkably tolerant of anaesthesia and handling. Fish have been observed feeding withintwohours of being anaesthetized. Morethanfive thousand fish were anaesthetized over the course of thisresearchand only a single fish died following the procedure.

2.6.2Gastric Evac.uatioD Triab

Fish were starVed for 5 days prior to theSlartof the experiment. Al4 p.m. 00 the day the experimeot started, fish were fed to apparent satiationwiththe diet lhey had been conswning for the pastthreemonths and the tank was immediately siphoned. Five randomly selected fish were removed from each tank at eight-hour intervals beginning fifteen minutes following the meal and every eight hours thereafter. The fifteen minute delay in taking thefirstsamplewastoensure that pellets had been completely swallowed and had moved into the stomach. Fish were killedwithan overdose of MS·222, and were immediately placed in plastic bags,sealed,thenfrozenina deep freezer with their heads slightly elevated.. This positionreducedany risk of ingested feed leaking out of the fish stomachs, since there is no valve between the esophagus and. lhe stomach (Verigina, 1974). There is a pyloric valve between the stomach and the intestine, so there was linle concern of losing stomach contents into the intestine. Fresh livers were generally friable so freezing made the samples more manageable.F~zingalso prevented leakage from stomachs, since the contents were often watery.

25

For analysis.fishwere set out on the laboratory bench for 10 minutes Wltil the layer of skin and musclecoveringthestomac!l thawed. The fishweteweighed and thestandard lengthwasrecorded. Attansverscventral incisionwasmade between the opercula.

followedbya medial, longitudinal incision to the analpore.The body wallwaspulled away, revealingthe bodycavity.Theliver,whichcoverstheesophagusand most of the stomach..wasremoved and wcigbcd.The stomach was removed by severing the esophagus next to the Stomach and the intestine immediately posterior to the pyloric valve. Since the stomachwasstillfrozenatthispoint, therewasDOconcern about contents leaking out. A small incision to the stomach wall allowed the contents to be simply squeezed out of the stomach onto a pre-weighed filter paper in a Buchner funnel.

The stomachwaH wasrinsed with distilled water and thefrozenstomach contents were thawed and distributed around the filterpaperusing distilled water.Thesepaperswere suction-filtered and ovendricdat60"cfortbrcc daysprior to rc-wcigbing (Dos Santos andJobling, 1988). The dJy weight of thestomach contents was calculated by sublraCtion.

2.6.3 Did ADalysis

Proximate analysesofdietarymoisture,protein,lipidandash~completedin triplicate for every diet and mean values arereportedin Table 2.1. The moisture contentwas determinedbydryingpre-weighed samplesinan oven set at105OC untilthedry weight wasconstant. Samples wert cooled in a dcssieatortominimizetheadbcIenccof water to Ute sample or sample dish. The protein content ofdried sampleswasdetermined using Ute Kjc1dah1 method (Tccator Digestion System 20,lOISdigester, Sweden; Tccator Kjel1cc::

System 1028 Distilling Unit, Sweden). Total nitrogenwasconverted tocrude:protein by multiplying by6.25on the assumption thattheproteininUte feed is approximately 16%

nitrogen. Thecrudelipid content of each diet was determined using a hexane-based Soxhlet lipid extraction apparatus (Tecator Soxtee System HT 1043 Extraction Unit, Sweden). The ash contentwasmeaswed by placing a pre-weighed crucibleand drieddiet

26

samplein a mufflefurnace(Thennolyne, Sybr'on Corporation,Dubuque,Iowa, USA) set at450

°c

overnight,coolingin adessieator and reweighing the cnx::ible and sample. The carbohydrate cootent of the dietswasestimatedby subttaeting thesumof the other nutrientsfrom100. Thegrossenergy of all dietswas calculated by multiplying the percent proteininthe diet by 5.6 kcallgram. the pen::entage of lipid by9.SkcaUgram and the percentage of carbohydrateby4.1 kcaUgram. Thesumof these values equals the grossdietary energy per100 grams.

2.7Calculations 2.7.1 Coaditioa ladex

The coodition index(Cn [elates the fish weight to its length. High CI values indicate a highweightperunit length, which is generally a favourable characteristic. The CI of each fishwascalculated using the following formula:

(2)

where W is wet weight(g) aDd L is standard length (em) (Goddard, 1996).

2.7.2 Speclflc Growth Rate

The~ificgrowth1m:(SGR)desc:ribesthe daily rate of growth as a percent ofbody weight. Itwascalculated according to the following fDmlula:

(3)

where Wttand Wtl arethe wet weights (g) oftbe individuals at day tl and tz (Goddard.

1996).

27

2.7.3 Feed CoavcnioD Ratio

The feed convenion ratio(FeR) describes how efficiently me feed is converted to body weight. tdeally, meFeRfor a dry diet is1.0or less.TheFeRfor this and subsequent trialswascalcwated as follows:

FeR· feed ingested(g) I(fW2·TW I (g), (4)

whereTWIandTW1are the sum of the weights of fish in a tank(g)at the beginning and eDdofthe feeding aia1(Goddard.1996).

2.7.4ProteiD Efficiency Ratio

Theproteinefficiency ratio(PER)givesanindieatiOQ of the weight gainpergram of protein ingested. ThePER for eachtank wascalculated as fonows (papoutsoglou et01.•

1987):

PER· 100 • (fWl· TW1

(g»

^I(feed ingested(g) •%protein), (5) whereTW1andTW2are the sum of the weights of fishinatank(g)at the beginning and end of the feeding triaL

2.7.5 Repato.omatic Index

The liver weight relative to the totalfishbody weight is indicated by me hepatosomatic index: (HSI). This index maybeanindirectdeterminant of the degree of lipid deposit in the liver. The HSI was calculated usina the following fonnula:

HSI- (L..../W) •100,

2.

(6)

whereLwislhe liver weight (g) and W is the weigh[ ofthewhole fish (g) (Stefanussen~t

at..1993).

2.7.6 CostofProductiOD

TheCOSIof productionwascalculated based only onfeed costsandgrowthtatcsinorder toevaluate lhe effectiveness of lhe diets. No factors such as pumping or heating costs were taken into account. The production costwascalculated using the fonnula:

$/kg produced· cost of food (Slkg) •FeR; (7)

where lbc cost of food (Slkg) was provided by the manufaclurer, and FCRwascalculated by Equation 4.

2.8 Statistical Ana/ys;s

2.8.1 FtediDg Trials

Treatment means ofthecoDditionindices,specificgrowthtaleS, and bepatosomaric indiceswerebased00measurements ofaU theindividualfishwithin each tank. Feed conversion ratios and protein efficiency ratios~calculatedODapertankbasis.

Analysis of variance (ANDYA) was dooe using SPSS software (SPSS, 1994; SPSS Inc., Chicago. Illinois)with0. -0.05. Insignificant factors were pooled and reanalysed.

Tukey's B multiple range testwasused todeterminethenatuleof significanl treatment differences. All percentagedatawere arcsine transfonned prior to analysis (Sakal and Rohlf, 1969). All means are reportedwith±standarderror.

29

2.8.2Gastric Evacuation Triab

Stomach contents at each sampling time were expressedingrams and as a percentage of body weight. Using regression analysis. the rate of evacuationwascalculated based on the percent body weight of food remaininginthe stomach and the weight of meal remaininginthe stomach. Regression coefficients were calculated using linear and logaritlunic models. All percentage data were arcsine transformed prior to analysis.

30

3.0 Results

3.1 StDcking Density! FHding ScheduleTri.' 3.1.1 ObservarioDJ

Aggressionwasnoted amongfishheld at 20 gIL. especiallythosefed once everytwo days. This aggressionwasgenerally initiated by thelargest fish in the tank, who would swim towards another with its mouth open. The aggressor generally did not bite the other fish, but simply hit themwith an open mouth between the pectoral fins and the anal pore.

No wounds or tomfinswere observed at any time. The aggressive behaviour was more prevalent during feeding butwasDOt limited to lhis time. When aggressionwasevident, onlythe aggressor consumed feed.

The feeding schedulewasa factor inthe rate offeed consumption.Fishfedtwiceperday generally waited forthepellettoreachthe bottom of thetankandmadeno responseto thepeUetforup to10seconds.Those fish fedoDCC daily and once every second daywere fed in the afternoonbasedonobservationsthat theyweremoreactive inthe afternoon and consumedmeals slightly faster. Thiswasimportant,becausewhen theywereslightly lethargic inthe morning,the feed sits onthetank bottom longerand likelybeco~less palatable. As a ruJe, when fish fed every second daywereoffered feed, they fed more activelythan fish fedmorefrequently.Fishfed once every second dayswamtothe water surface wheneverthe technician approached, regardless of wbether foodwasoffered or not.This behaviourwasless evidentinfisb fed once daily andwasnever notedinfish fed twice daily.

31